A traffic state identification method based on improved cloud model

Traffic state identification is the foundation of making scientific traffic organization scheme and traffic management strategy. The multiple source sensor data fusion is a kind of effective method to realize the traffic state identification. In order to improve the identification accuracy of traffic state, a decision fusion model based on synthesized cloud and rough set theory is presented. In the method, several experts are called to given the threshold value of parameters from different detectors, and these parameters are used to generated traffic state clouds, then synthesized cloud is proposed to fusion these traffic state clouds, and it is the first level fusion. In second level fusion, importance degree of attribution of rough set is proposed to compute the weight of each expert to make sure the best expert has a highest weight. The proposed fusion method can achieve the traffic state identification and the rate of state recognition is improved obviously.